Lyrebird API
Date | 2017-04-24 20:00 |
From | Rory Walsh |
Subject | Lyrebird API |
Csound mailing list Csound@listserv.heanet.ie https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here |
Date | 2017-04-24 20:24 |
From | Guillermo Senna |
Subject | Re: Lyrebird API |
That is great! A little rough yet, but I can't wait to hear all of your voices saying that SuperCollider is a much better language than Csound! On 24/04/17 16:00, Rory Walsh wrote: > I can't tell if this is a joke or not? > > https://lyrebird.ai/ > > Click the demo tab.. > > Csound mailing list > Csound@listserv.heanet.ie > https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND > Send bugs reports to > https://github.com/csound/csound/issues > Discussions of bugs and features can be posted here > Csound mailing list Csound@listserv.heanet.ie https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here |
Date | 2017-04-24 20:32 |
From | Oeyvind Brandtsegg |
Subject | Re: Lyrebird API |
Wasn't there a similar thing announced by Adobe a while back too? It (the demos) sounds plausible enough in light of the current technology, but a bit strange that they make the demos sound like low bitrate voice compression... see also 2017-04-24 12:00 GMT-07:00 Rory Walsh <rorywalsh@ear.ie>: Csound mailing list Csound@listserv.heanet.ie https://listserv.heanet.ie/ Oeyvind Brandtsegg Professor of Music Technology NTNU 7491 Trondheim Norway Cell: +47 92 203 205 http://www.partikkelaudio.com/ http://crossadaptive.hf.ntnu.no http://gdsp.hf.ntnu.no/ http://soundcloud.com/brandtsegg http://flyndresang.no/ http://soundcloud.com/t-emp |
Date | 2017-04-24 20:42 |
From | Matt Berlin |
Subject | Re: Lyrebird API |
They make a big deal of the ethics involved regarding possible impersonation, but the inflection changes are very discernible, as are the generated samples.
Csound mailing list
Csound@listserv.heanet.ie
https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
Send bugs reports to
https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here
I see the most likely use being matching a Bot-voice with a user's regional dialect (call centers, and burgeoning bot-salesmen services). - Matt On Mon, Apr 24, 2017 at 3:32 PM, Oeyvind Brandtsegg <oyvind.brandtsegg@ntnu.no> wrote:
|
Date | 2017-04-24 20:48 |
From | Richard |
Subject | Re: Lyrebird API |
Sounds like something that csound could do with some PVS opcodes.. Richard On 24/04/17 21:24, Guillermo Senna wrote: > That is great! A little rough yet, but I can't wait to hear all of your > voices saying that SuperCollider is a much better language than Csound! > > > On 24/04/17 16:00, Rory Walsh wrote: >> I can't tell if this is a joke or not? >> >> https://lyrebird.ai/ >> >> Click the demo tab.. >> >> Csound mailing list >> Csound@listserv.heanet.ie >> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND >> Send bugs reports to >> https://github.com/csound/csound/issues >> Discussions of bugs and features can be posted here >> > Csound mailing list > Csound@listserv.heanet.ie > https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND > Send bugs reports to > https://github.com/csound/csound/issues > Discussions of bugs and features can be posted here Csound mailing list Csound@listserv.heanet.ie https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here |
Date | 2017-04-24 22:47 |
From | Rory Walsh |
Subject | Re: Lyrebird API |
That is great! A little rough yet, but I can't wait to hear all of your Ha! Evil genius!
|
Date | 2017-04-26 00:16 |
From | "Dr. Richard Boulanger" |
Subject | Re: Lyrebird API |
WOW. When will we have the Lyrebird opcode family in Csound? I can wait! -dB _____________________________________________ Dr. Richard Boulanger Professor of Electronic Production and Design Professional Writing and Music Technology Division Berklee College of Music ______________________________________________ President of Boulanger Labs - http://boulangerlabs.com Author & Editor of The Csound Book - http://mitpress.mit.edu/books/csound-book Author & Editor of The Audio Programming Book - http://mitpress.mit.edu/books/audio-programming-book ______________________________________________ about: http://www.boulangerlabs.com/#about about: http://www.csounds.com/community/developers/dr-richard-boulanger/ music: http://www.csounds.com/community/developers/dr-richard-boulanger/dr-richard-boulanger-music/ ______________________________________________ email: rboulanger@berklee.edu facebook: https://www.facebook.com/richard.boulanger.58 On Mon, Apr 24, 2017 at 3:00 PM, Rory Walsh <rorywalsh@ear.ie> wrote: Csound mailing list Csound@listserv.heanet.ie https://listserv.heanet.ie/ |
Date | 2017-04-29 19:04 |
From | Rory Walsh |
Subject | Re: Lyrebird API |
How about this: On 26 Apr 2017 12:16 am, "Dr. Richard Boulanger" <rboulanger@berklee.edu> wrote:
|
Date | 2017-04-29 19:16 |
From | Guillermo Senna |
Subject | Re: Lyrebird API |
Are those all synthetic voices?? Having that kind of tech in Opcode form is going to be so much fun. On 29/04/17 15:04, Rory Walsh wrote: > How about this: > http://www.dtic.upf.edu/~mblaauw/IS2017_NPSS/ > > On 26 Apr 2017 12:16 am, "Dr. Richard Boulanger" |
Date | 2017-04-29 20:19 |
From | Michael Gogins |
Subject | Re: Lyrebird API |
I have already created an issue in GitHub for a Csound opcode or opcodes to do singing synthesis. I have researched this topic fairly extensively (https://github.com/csound/csound/issues/658). It is not ready to go. Here are the reasons: 1. There is money in the singing synthesis business, therefore when some technology begins to work it is bought or otherwise made proprietary, such that the code is not available. 2. Most of the interest in singing synthesis comes from "otaku" types in Japan and China. There is much more work on singing synthesis in Japanese and Chinese than there is on singing synthesis in English. That said, not only the Lyrebird examples but also other examples e.g. from Sinsy and Kanru Hua's Moresampler are musically useful and would be great to have in Csound. Thus, I am continuing to track the state of the art and the state of the open source code base in this field. I do not know what might come of the Lyrebird project, but it sounds like a client/server approach where a compute farm uses deep learning and other, more traditional techniques to perform synthesis and return results in audio clips. It remains to be seen if my idea is correct and whether the approach could be used in Csound. Certainly the audio results are impressive and would be very useful. The approach I am currently pursuing (when I have the time) is to adapt the phonetics inputs for the Japanese-speaking version of Sinsy to use English phonetics instead of Japanese phonetics, by creating a map from English phonetic spelling to Japanese phonetic spelling. In all likelihood the adaptation would be imperfect, but this is currently the only approach based on publicly available code that has a chance of working well enough to be used in Csound, as far as I know. I will keep on looking.... Best, Mike ----------------------------------------------------- Michael Gogins Irreducible Productions http://michaelgogins.tumblr.com Michael dot Gogins at gmail dot com On Sat, Apr 29, 2017 at 2:16 PM, Guillermo Senna |
Date | 2017-04-29 22:23 |
From | Rory Walsh |
Subject | Re: Lyrebird API |
Pdf here: On 29 Apr 2017 8:20 pm, "Michael Gogins" <michael.gogins@gmail.com> wrote: I have already created an issue in GitHub for a Csound opcode or |
Date | 2017-04-30 00:00 |
From | Richard |
Subject | Re: Lyrebird API |
Quite impressive... Richard On 29/04/17 20:04, Rory Walsh wrote:
|
Date | 2017-04-30 00:22 |
From | Michael Gogins |
Subject | Re: Lyrebird API |
It's impressive all right, but I would be surprised if there were anything open source about it. The Universitat Pompeu Fabra is the same outfit where Xavier Serra developed the technology that eventually become the Vocaloid, which is definitely not open source. The UPF does make a lot of very good source code publicly available. E.g. the UPF's Essentia library is Affero GPL. This is a strong copyleft or "free software" license, that is, IT IS NOT REALLY OPEN SOURCE and IT IS NOT COMPATIBLE WITH THE CSOUND LICENSE. Too bad! Just to remind you all, "free software" means source code must be published AND any derivative product must also be "free software," and "open source" (which is what Csound's Lesser GPL v2 is) means only that source code must be published and can be used to make any kind of derivative product. You cannot usually use code with "free software" licenses in to make "open source" products. Regards, Mike ----------------------------------------------------- Michael Gogins Irreducible Productions http://michaelgogins.tumblr.com Michael dot Gogins at gmail dot com On Sat, Apr 29, 2017 at 7:00 PM, Richard |